JP4326999B2 - Image processing system - Google Patents

Description

  The present invention relates to an image processing system suitable for detecting foreign matters adhering to the surface of various window glasses.

  Droplets such as raindrops and fogging and dust adhering to the surface of various types of window glass such as glass used in vehicles, ships, aircraft, etc. or window glass of general buildings, etc., reduce visibility through the window glass. . 2. Description of the Related Art Conventionally, a method is known in which foreign matter attached to a window glass is detected by an image processing system and then the foreign matter attached to the window glass is removed by various means.

  In Patent Document 1, the window glass is irradiated with two light sources, the reflected light from the window glass is detected by the image sensor, and raindrops attached to the outer surface of the window glass and the inside of the window glass are detected from the intensity of the reflected light. A method for detecting haze adhering to the surface is described. In Patent Document 2, in order to simplify the configuration of the image processing system, the window glass is irradiated using one light source and a lens, the reflected light from the window glass is detected by an image sensor, and the outside of the window glass is detected. A method for detecting foreign matter adhering to the surface is described. In Patent Document 3, a filter in front of one in-vehicle camera is mechanically put in and out using a motor to acquire an image under a plurality of conditions, and foreign matter attached to the outer surface of the window glass is detected. A method is described.

Japanese Patent Publication No. 07-89099 JP-T-2001-516670 Japanese Patent Laid-Open No. 11-160777

However, since the above prior art detects the reflected light of the two light sources, it is necessary to apply the reflected light to an independent area of the image sensor, and an optical system for that is essential, and it is difficult to simplify the apparatus. There is also a problem that the influence of external light such as sunlight cannot be removed. In addition, the method of detecting the reflected light from one light source is difficult to distinguish between the reflected light from the outside of the window glass or the reflected light from the inside, and the host vehicle has the lens focused on the window glass. There has been a problem that it is difficult to simultaneously detect the white line ahead and the preceding vehicle.
Further, in the above-described conventional technology, it is not possible to capture images at the same time under a plurality of conditions, or to capture images of a plurality of conditions continuously in time. In the future, in order to acquire a large amount of information in a short time and to perform high-speed image processing, it is expected that the imaging device and the processing device will increase in speed. It is difficult to deal with processing. In addition, products mounted on automobiles are required to have high quality that can withstand a wide range of temperature changes and humidity changes, and those having mechanical moving parts as in the above-mentioned prior art are preferable in terms of reliability. Absent.

  The present invention has been made in view of the above-mentioned problems, and without using a mechanical movable part, the image of a plurality of conditions is captured using a single camera, and foreign matter adhering to the window glass is accurately detected. An object is to provide an image processing system capable of detection.

  The purpose is to provide a light source that emits light toward the glass, an image sensor that captures reflected light reflected by a foreign object attached to the glass, and image processing that processes the image captured by the image sensor. This is achieved by providing an apparatus and providing an optical filter that transmits a specific wavelength of reflected light between the glass and the imaging device to constitute an image processing system. As the optical filter, for example, a filter having a characteristic that transmits light in a wavelength region in which the transmittance of the window glass is equal to or higher than a predetermined value and the sensitivity of the imaging element is high is used. Alternatively, a wavelength that cuts the wavelength in the visible region and transmits light in a wavelength region in which the imaging element has sensitivity on the longer wavelength side than the visible region is used. In the optical filter, the optical characteristics of a single filter may differ depending on the area of the filter, or the characteristics of a lens or lens cover provided in front of the image sensor may be the same as the above-described filter.

  According to the present invention, it is possible to easily separate and detect raindrops adhering to the outside of the window glass and cloudiness adhering to the inside without being affected by sunlight, ambient light, and day and night by a simplified configuration. it can. In addition, since it is possible to capture images of a plurality of conditions using a single imaging device without providing a movable part, it is possible to realize a plurality of functions based on information from a single imaging device, and further downsize the imaging device. Also contributes.

  Embodiments of the present invention will be described below with reference to the drawings. Here, an embodiment in the case where the present invention is mainly applied to an automobile will be described.

  FIG. 1 shows a schematic diagram of an example of an image processing system according to the present invention. FIG. 1A is a side sectional view, and FIG. 1B is a front view seen from the window glass side of an automobile.

  The image processing system according to this embodiment includes an image processing device 1, a first light source 2, and a second light source 3, and the light emission timings of the first light source 2 and the second light source 3 are controlled by the image processing device 1. Has been. The image processing apparatus 1 is installed on the indoor side of an automobile so as to face a window glass (front glass) 4. The first light source 2 is disposed on the side of the image processing apparatus 1 and emits light toward the window glass 4. On the other hand, the second light source 3 irradiates the window glass 4 from below.

  The first light source 2 emits light toward the window glass 4 using a drive circuit provided in the image processing apparatus 1. The light emitted from the first light source 2 passes through the window glass 4 when the window glass 4 is clean. Further, the light reflected by the window glass 4 is directed downward and does not enter the image processing apparatus 1. However, when raindrops 8 are attached to the outside of the window glass 4, the light emitted from the first light source 2 is reflected by the raindrops 8. Next to the first light source 2, the lens 10 of the image processing apparatus 1 is arranged. For this reason, the light reflected by the raindrop 8 passes through the lens 10 of the image processing apparatus 1 and enters the image sensor 30 via the optical filter 20 having predetermined optical characteristics. When raindrops 8 are not attached to the outside of the window glass 4, the light emitted from the first light source 2 passes through the window glass 4 and travels straight and does not enter the image sensor 30. FIG. 2A shows an example of an image photographed by the image sensor 30 when a large amount of raindrops adheres to the outside of the window glass 4. Raindrops are photographed as a shape close to a circle. In particular, if the window glass 4 is provided with water repellency, raindrops can be photographed in a more circular shape.

  Next, the second light source 3 is disposed on a dashboard 6 in the automobile, which is below the image processing apparatus 1. The second light source 3 emits light toward the window glass 4 using a drive circuit provided in the image processing apparatus 1. The light emitted from the second light source 3 is totally reflected inside the window glass 4, and the reflected light enters the image sensor 30 through the lens 10 of the image processing apparatus 1. However, when the cloud 12 is attached to the inside of the window glass 4, the light emitted from the second light source 3 is scattered by the cloud 12, and a part of the light passes through the lens 10 of the image processing apparatus 1 and is optical. The light enters the image sensor 30 through the filter 20. An example of an image photographed by the image sensor 30 when the upper portion of the window glass 4 is cloudy is shown in FIG. Cloudiness hinders forward visibility.

  In the embodiment of FIG. 1, a control signal line is required from the image processing apparatus 1 to the second light source 3 arranged on the dashboard 6, and the control signal line becomes long. FIG. 3 shows another embodiment of the present invention. In FIG. 3, in order to make the image processing system of the present invention compact, the image processing apparatus 1, the first light source 2, and the second light source 3 are incorporated in the same case 5 to constitute an image processing system. Yes.

  Furthermore, another embodiment is shown in FIG. In the present embodiment, the second light source 3 is also disposed on the side of the image processing apparatus 1, and the mirror 60 is disposed at the position of the case 5 where the light from the second light source 3 is reflected by the window glass 4. . In the present embodiment, the light emitted from the second light source 3 is reflected by the window glass 4, is incident on the mirror 60, is reflected there, and is further reflected by the window glass 4 to the image sensor 30 of the image processing apparatus 1. Incident. The mirror 60 may be installed on the dashboard instead of being installed in the case 5. The light from the second light source 3 is reflected by the mirror 60 and enters the lens 10 of the image processing apparatus 1, but the light from the first light source 2 is reflected by the mirror 60 and the lens of the image processing apparatus 1. The arrangement of the first light source 2, the second light source 3, and the mirror 60 is determined so as not to be incident on the light source 10.

  In the image processing system shown in FIG. 3, in order to cause the light emitted from the second light source 3 to be totally reflected inside the window glass 4 and to be incident on the image pickup device 30, the second light source 3 is placed on the window glass 4. It needs to be placed nearby. If the image processing apparatus 1, the first light source 2, and the second light source 3 are surrounded by the case 5, even if fogging occurs in other parts of the window glass 4, the window glass 4 near the image processing system is fogged. May be prevented. Therefore, by providing the air flow passage 11 in the case 5, when fogging occurs in other parts of the window glass 4, the window glass 4 in the vicinity of the image processing system is also fogged.

  1, 3, and 4, the image processing device 1, the first light source 2, and the second light source 3 are used. The image processing apparatus 1 and the second light source 3 may be used.

Visible light or infrared light may be used as the emission color of the first light source 2 and the second light source 3 used in the embodiment of the present invention. If there is a restriction on the wavelength of a light source that can be mounted on an automobile, a specific wavelength may be selected from wavelengths having good transmittance of the window glass 4. For example, Figure 5 transmission 32 is a predetermined value of the window glass 4 shown in (a) (FIGS. 5 (a) in 70%) or more with a sensitivity 31 is high wavelength of the imaging element 30 (wavelength _{f a} following. Specifically The wavelength in the visible light region within 400 to 600 nm may be selected. Alternatively, in order not to dazzle oncoming vehicles and pedestrians, the wavelength f _b (for example, an infrared light region in the range of 800 to 1000 nm) whose wavelength is longer than visible light and the light receiving sensitivity 31 of the image sensor 30 extends. Wavelength). The characteristics of the optical filter 20 may be selected corresponding to the wavelengths of the first light source 2 and the second light source 3 selected in this way. In other words, when a light source having a wavelength in the latter infrared region is selected, the wavelength in the visible region is cut, and light in the wavelength region where the imaging device has sensitivity on the longer wavelength side than the visible region is transmitted. The optical filter may be used.

A configuration example of the optical filter 20 is shown in FIG. FIG. 6A shows an example in which the light incident surface of the image sensor 30 is covered with a band-pass filter 21 that transmits a specific wavelength. This is an example in the case of detecting only raindrops and cloudiness. Light of a specific wavelength is irradiated from the light source 2 or light source 3 and only the light of that wavelength can be received, so that it is not affected by external light. FIG. 6B is an example in which only a partial region is covered with the band-pass filter 21 on one side on the same side as FIG. Although FIG. 6A can receive only light of a specific wavelength with the image sensor 30, FIG. 6B can receive not only the specific wavelength but also other wavelengths of light, so that an image outside the vehicle is captured through the window glass 4. Can be processed. Furthermore, when the infrared light transmittance of the window glass 4 is large, as shown in FIG. 6C, in order to reduce the influence of sunlight, which is external light, as shown in FIG. 6A and FIG. An infrared light cut filter 22 may be disposed on the light incident side of the band pass filter 21. Further, as shown in FIG. 6D, one optical filter 20 may have an infrared light cut filter region and a band pass filter region. Here, an example of the transmittance of the infrared light cut filter and the band pass filter (wavelength f _c ) is shown in FIG. The band-pass filter, for example, in the central transmission wavelength f _c is 850 nm, width or half width of wavelength at which the transmittance becomes half can be used to pass wavelengths of ± 20 nm.

  As shown in FIG. 6D, when light that has passed through the infrared light cut filter is incident on a part of the image sensor 30 and light that has passed through the band pass filter is incident on another area, It is possible to detect not only raindrops and cloudiness adhering to the window glass 4, but also other objects such as vehicles and lanes, an example of which is shown in FIG. This embodiment shows a case where the headlight orientation control function and the raindrop detection function are simultaneously established by a single image pickup apparatus. As shown in FIG. 7A, the in-vehicle image processing apparatus is installed in front of the vehicle, captures a front scene through the windshield, and captures an image as shown in FIG. The light distribution control function performs image processing from the image of FIG. 7B, and extracts the taillight 230 of the preceding vehicle 210 and the headlight 240 of the oncoming vehicle 220 as shown in FIG. 7C. And it is the function which measures the position of a preceding vehicle and an oncoming vehicle from the image of FIG.7 (c), and controls the orientation of the headlight of the own vehicle according to the position of a preceding vehicle and an oncoming vehicle. In addition, 250 is a horizon.

  Therefore, as shown in FIG. 7B, the imaging range in which the headlights and taillights of a distant vehicle and the headlights and taillights of a nearby vehicle are reflected is an image processing range necessary for realizing the orientation control function. . Here, when recognizing a preceding vehicle, it is necessary to determine the presence or absence of a preceding vehicle with the taillight of the preceding vehicle, but the amount of light is small compared to the headlight of the oncoming vehicle, and there are many disturbing lights such as street lights Therefore, it is difficult to detect the taillight only from the luminance, and it is necessary to recognize the red color of the taillight.

  However, when a color image sensor is used as the image sensor 30 for recognizing red, the color image sensor has sensitivity also in the infrared region. Therefore, when an image is captured only by the image sensor, the entire image data obtained is Therefore, it is difficult to extract the red portion of the taillight. Therefore, for example, by installing a filter having a characteristic of cutting infrared light as shown in FIG. 8 at the position of the optical filter 20 in FIG. 7A or in front of the lens 10, red and other colors can be obtained. It becomes easy to make a distinction. Thereby, since the light of the other color used as disturbance can be removed, the detection accuracy of a taillight can be improved. Here, in the filter having the characteristic of cutting infrared light as shown in FIG. 8, the wavelength x to be cut, the method of decreasing the transmittance, etc., are used for each application, the performance of the imaging device, and the window glass 4 shown in FIG. Select one with characteristics suitable for transmittance. Typically, the wavelength x in FIG. 8 may be about 600 nm.

  Hereinafter, the wavelength of the light source 2 is an infrared wavelength, for example, 850 nm. Various light sources such as a light emitting diode (LED) and a semiconductor laser can be used as the light source 2 for irradiating light of infrared wavelength. Here, as shown in FIG. 9A, for example, the light source 2 that irradiates infrared light is installed at an angle at which the light reflected by the raindrops 8 attached to the window glass 4 is incident on the image sensor 30. The infrared light 43 emitted toward the glass 4 may be radial light or parallel light.

  Here, if it is going to detect the light of the infrared wavelength reflected in the water droplet as it is, the light source 2 which irradiates the light of the infrared wavelength brightens the light irradiated rather than the disturbance light which has enormous light quantity, such as sunlight. There is a problem of having to.

  Therefore, for example, as shown in FIG. 9B, a filter that cuts light having a wavelength shorter than the emission wavelength of the light source 2 or a transmittance peak as shown in FIG. A band-pass filter substantially matched with the emission wavelength is installed at the position of the optical filter 20 in FIG. Thereby, light other than the light emission wavelength of the required light source 2 can be removed, and the light quantity of the detected light source 2 can be made relatively large. Here, in FIG. 9B, the wavelength y to be cut and how to reduce the transmittance, and in FIG. 9C, the characteristics such as the transmittance peak wavelength z, the transmittance, and the full width at half maximum of the bandpass filter are used for each purpose. Or a device suitable for the performance of the imaging device may be selected. FIG. 9D shows an example of a raindrop image captured by the image sensor 30 using such a filter.

  FIG. 10 is a diagram illustrating an example of a filter having different optical characteristics depending on regions. For example, the light distribution control is performed at the top 2/3 of the screen and the image processing is performed so that raindrop detection is performed at the bottom 1/3 of the screen. A region having a cut characteristic (hereinafter referred to as “infrared cut region”) 51a and a region having a band pass characteristic having a transmittance peak in the infrared region (hereinafter referred to as “infrared band pass region”) 52a. Create separately. For example, the configuration is as shown in FIG. By using one such filter, only one imaging unit, which has conventionally been required for the number of functions, can be used, and a significant reduction in size and cost can be achieved.

  Further, by using one filter, reflection between filters is reduced, and disturbance light that is disadvantageous for image processing can be reduced. In addition, since there is no mechanical moving part and the filter is fixed, it is possible to capture a plurality of videos continuously even if the speed of capturing an image is increased.

  Here, as shown in FIGS. 10B to 10D, various methods can be used for dividing the filter area. However, as shown in FIG. 7, the headlight of the oncoming vehicle and the taillight of the preceding vehicle are often reflected mainly at the top of the screen, and the road surface near the vehicle is usually reflected at the bottom of the screen. In control, information at the top of the screen is important, and information at the bottom of the screen is not very important. Therefore, in order to achieve both the light distribution control function and the raindrop detection function, the lower part of the screen is used as an infrared bandpass region for raindrop detection as described above, and the remaining region is used as an infrared cut region for light distribution control. The configuration shown in FIG. 10A is preferable.

  FIG. 11 shows an imaging apparatus having the optical filter 20 shown in FIG. FIG. 12 shows an image acquired when the optical filter 20 is attached to the imaging apparatus as shown in FIG. Here, due to the characteristics of the lens 10, the top and bottom are reversed between the actual scene to be imaged and the image condensed on the image sensor 30. Therefore, in order to make the lower part of the screen an infrared bandpass region for raindrop detection, the top side of the optical filter 20 is an infrared bandpass region 52a. An example of an image of a vehicle and raindrops taken using this imaging apparatus is shown in FIG.

  The image pickup unit is configured as described above, and the image processing unit uses the image (hereinafter referred to as “infrared band pass region on the image pickup device”) where light transmitted through the infrared band pass region is collected as raindrops. By performing light distribution control using an image (hereinafter referred to as “infrared cut area on the imaging device”) where light that has passed through the infrared cut area is collected by detection, with a single filter, A raindrop detection function and a light distribution control function can both be achieved without providing a mechanical movable part.

  Here, as shown in FIG. 11, the imaging unit may have a configuration in which there is a gap between the optical filter 20 and the imaging device 30, but the configuration in which the optical filter 20 is closely attached to the imaging device 30 is optical. It becomes easy to make the boundary of the region on the filter 20 coincide with the boundary of the region on the image sensor 30.

  The infrared cut region and the infrared band pass region on the image sensor 30 can be determined by the pixel number of the image sensor 30. That is, in the example of FIG. 10A, the upper end of the image sensor 30 to the Nth pixel, for example, is an infrared bandpass region, and the lower end of the image sensor 30 to, for example, the Mth pixel is an infrared cut region. You can also

  In addition, the amount of light that passes through the infrared bandpass region and enters the image sensor 30 is smaller than the amount of light that enters through the infrared cut region. For this reason, if the shutter speed of the imaging device is set to a shutter speed that does not saturate the imaging device 30 in the infrared cut area depending on the traveling environment and the light beam state, the image in the infrared light bandpass area becomes dark due to insufficient exposure. For this reason, the taillights and headlights of vehicles in the vicinity of the vehicle may not be visible. Therefore, as shown in FIG. 13A, by tilting the image processing apparatus 1 including the imaging unit and the image processing unit forward from the horizontal direction, as shown in FIG. The headlight, the taillight of the nearby vehicle, and the headlight are reflected in the infrared cut region, and the light distribution control function can be established. Here, the forward tilt angle depends on the height of the vehicle, the size and arrangement of the lens 10 and the image sensor 30, and is preferably about 7 °. Further, the same effect can be obtained by a configuration in which only the imaging unit is tilted forward, or a configuration in which the angle of view of the lens is expanded downward to widen the imaging range.

  As another embodiment, a case where the lane detection function and the raindrop detection function are simultaneously established will be described. First, in order to realize the raindrop detection function, as described above, the upper part of the filter may be an infrared bandpass region as shown in FIG.

  On the other hand, in order to realize the lane detection function, it is particularly important to detect a lane in the vicinity of the own vehicle. Because, as shown in FIG. 14, the lane has not only the straight line 81 but also the curve 82, the distance of the lane in the screen increases as the distance increases. Conversely, the lane in the vicinity of the host vehicle travels in the lane. This is because there is a feature that the amount of movement in the screen is small.

  Therefore, if the upper part of the filter is the infrared bandpass region as described above, the image information in the vicinity of the host vehicle, which is particularly important for the lane detection function, appears on the lower side of the screen. However, since the image information of the lane is the wavelength in the visible light region, it is attenuated by the above bandpass characteristic and the amount of transmitted light is reduced. It becomes difficult.

  In addition to the above, when trying to detect a curve from the lane shape, it is necessary to detect not only the vicinity of the host vehicle but also a distant lane. It is also undesirable to install an area below the filter and use this area to detect raindrops.

  Therefore, when the image processing apparatus 1 is tilted forward by, for example, 7 ° as shown in FIG. 13A, all the lanes from a distance to the vicinity of the host vehicle are shown in the infrared bandpass region as shown in FIG. The image can be projected outside, and the raindrop detection function and the lane detection function can be realized simultaneously. Here, only the imaging unit may be tilted. Also, widening the shooting range using a lens with a wide angle of view makes it possible to project from a distance to the vicinity of the host vehicle while securing a region necessary for raindrop detection.

  Normally, when the camera is tilted down, the hood of the vehicle appears, and sunlight reflected by the hood of the vehicle and the taillight of the preceding vehicle become disturbances, which is disadvantageous for lane detection and light distribution control. It becomes a condition. However, for example, when a filter as shown in FIG. 10A is used, disturbance light can be removed and the area can be assigned to an area where image processing for raindrop detection is performed due to the characteristics of the bandpass. By using the filter, it is possible to perform image processing in an area that has been unsuitable for image processing. In addition, an infrared light source having the same emission wavelength as the bandpass transmittance is provided, and by irradiating it, it is possible to detect an obstacle on the road and to specify the position of the preceding vehicle.

  In these embodiments, the best configuration has been described using a configuration in which a filter is provided between the lens 10 and the imaging device 30, but a lens or a lens cover attached to the front surface of the lens is similar to the above-described filter. The same effect can be obtained even with a configuration having optical characteristics.

  With reference to FIG. 15, a method for detecting raindrops adhering to the outside of the window glass 4 and fogging adhering to the inside of the window glass 4 using the image processing system of the present invention will be described. FIG. 15A is a diagram showing a light emitting circuit configuration example of the light source 2 and the light source 3, and FIG. 15B is a diagram showing an example of light emission timing and raindrop and cloudy detection timing in synchronization with the vertical synchronizing signal of the image. is there. In order to emit the light source 2, the image processing apparatus 1 controls the base voltage of the transistor 9-1 to pass a predetermined current through the light source 2, and to emit the light source 3, the base of the transistor 9-2 is used. A predetermined current is passed through the light source 3 by controlling the voltage. First, when detecting raindrops adhering to the outside of the window glass 4, the image processing apparatus 1 causes the first light source 2 to emit light and irradiate the window glass 4. When raindrops are attached to the outside of the window glass 4, the light emitted from the first light source 2 is scattered by the raindrops and part of the light enters the image sensor 30 as reflected light. The image sensor 30 generates an image signal using incident light as an image. The image signal is input to an image processor (not shown). And the raindrop adhering to the outer side of the window glass 4 is detectable by processing the image signal with an image processor. Here, the image processor is arranged in the image processing apparatus 1, but it may be arranged as an apparatus different from the image processing apparatus 1.

  An example of the flow of the raindrop detection program by the image processor is shown in FIG. In the process of FIG. 16A, the image processor inputs an image signal when the light source 2 is turned on from the image sensor 30 in Step 1. In step 2, edge detection processing is performed on the image signal using, for example, a known Laplacian filter. By this edge detection processing, an image in which the boundary between the raindrop area and the non-raindrop area is emphasized can be created.

  Next, a circle detection process is executed in step 3 for the edge image created in step 2. As the circle detection process, a generalized Hough transform may be performed. After performing circle detection in step 3, the number of circles detected in step 4 is counted. Next, in step 5, the number of circles is converted into rainfall, the rainfall is obtained, and the process is terminated.

  If the water repellent treatment is performed on the outer surface of the window glass 4, raindrops become circular, which is convenient for the treatment. Further, if the focal point of the lens 10 is focused not on the position of the window glass 4, but on the far side, the raindrops are blurred and can be picked up as a circle, so that the raindrops can be detected more easily. If the raindrops can be individually distinguished as a circle, the amount of rain per predetermined time can be calculated by the image processing apparatus 1.

  Further, as another processing example, a method shown in FIG. In FIG. 16B, the image processor inputs image signals when the light source 2 is turned on and off in Step 21 from the image sensor 30. Then, in step 22, a luminance difference is calculated for these image signals. Next, for example, an average value is obtained with respect to the luminance difference created in step 22, and a predetermined threshold value and magnitude determination process is executed in step 23. If the brightness difference is larger than the threshold value, it is determined in step 24 that rain and dirt have been detected, the brightness difference created in step 22 is converted to rain in step 25, and the process is terminated. If the luminance difference is smaller than the threshold value, the process ends. According to this method, for example, even when raindrops cannot be captured as a circle like drizzle, the amount of rain can be detected.

  The processes of FIGS. 16A and 16B may be used independently or in combination. By combining, it is possible to measure the rainfall with higher accuracy. A program for the processing flow shown in FIGS. 16A and 16B may be stored in the storage unit in the image processing apparatus 1.

  The image processing apparatus 1 can control the wiper or the washer based on the rainfall calculated according to FIGS. An example of the control configuration of the wiper and washer is shown in FIG. FIG. 17 includes a wiper / washer control unit 100, a wiper 160, a motor 150 that moves the wiper 160, and a washer control unit 170.

  An example of a processing flow for controlling the wiper 160 by the wiper / washer control unit 100 is shown in FIG. In FIG. 18, first, at step 100, it is determined whether or not the position of the shift lever is in the parking range (P range). If it is within the P range, the operation of the wiper 160 is prohibited (step 101). Thereby, the malfunctioning of the wiper 160 at the time of a car wash can be prevented. If it is not in the P range, the routine proceeds to step 110 where the wiper 160 is activated.

An example of a method for starting the wiper 160 in step 110 is shown in FIG. In FIG. 19, time is plotted on the horizontal axis and rainfall is plotted on the vertical axis, and the wiper 160 is activated when the rainfall exceeds the threshold. In step 111, it is determined whether or not the vehicle speed V is 0. If V = 0, the rain determination threshold value for starting the wiper 160 is increased in step 112. In step 116, the waiting time is set to a predetermined value. If the vehicle speed determined in step 111 is not 0, the rain determination threshold is changed to a predetermined value or according to the vehicle speed. In step 113, the threshold value is changed according to the vehicle speed. Then, after the wiper 160 is activated, a waiting time until the wiper 160 is activated next is calculated. In step 114, the waiting time is inversely proportional to the vehicle speed V (k is a proportional coefficient). Not only the vehicle speed V but also the rain amount and rain particle size may be determined according to the following equation. Here, k1, k2, and k3 are proportional coefficients.
Waiting time = k1 / rainfall + k2 * particle size + k3 / vehicle speed

  By this waiting time, it is possible to prevent the wiper 160 from being activated immediately upon detection of a rain film or rain streaks caused by the wiper blade wiping away raindrops. Next, in step 115, only the waiting time is waited, and in step 102, the rainfall is compared with the threshold value for starting the wiper 160. If the rainfall is greater than the threshold, the wiper 160 is activated in step 103 and the process is terminated. If the rainfall is smaller than the threshold, the process is terminated.

  An example of another wiper activation method is shown in FIG. Step 120 shown in FIG. 20 can be used in place of step 110 in FIG. 18, and controls intermittent time (interval) used as wiper control. In FIG. 20, k4, k5, and k6 are proportional coefficients. For example, the interval is controlled using the measured rainfall and rain particle size. When there is a lot of rainfall or when the particle size is small, the interval may be shortened. Conversely, when the rainfall is small or the particle size is large, the interval may be lengthened. At this time, the interval may be changed according to the traveling speed V of the host vehicle. For example, when the traveling speed is high, raindrops flow on the window glass 4 due to the wind speed, so the interval may be lengthened.

  In the embodiment of raindrops, the particle size is measured. However, the better the water repellency of the window glass 4, the closer the raindrop is to a circle, and the easier it is to measure the particle size. The degree of water repellency of the window glass 4 can be determined from the image input in FIG. 16 using the shape of raindrops, reflection luminance, and the like. Further, if the image after the wiper 160 is activated is input to detect the amount and shape of the raindrops (striped raindrops), the damage to the blades of the wiper 160 can be detected. The detected degree of water repellency and damage to the blade are detected by detecting the current flowing through the motor 150 in FIG. 17 to estimate the torque of the motor 150. If this torque increases, the water repellency deteriorates or the blade is damaged. It can also be determined that damage has occurred. Then, by alerting the driver of the detected water repellency and blade damage, it is possible to always ensure a good field of view.

  If the wiper 160 is further detected at the same position after the wiper 160 is activated due to rain or dirt detected in step 24 described with reference to FIG. 16B, it is determined that dirt is attached to the window glass 4, and washer control is performed. You may perform control which controls the part 170 and applies water to the window glass 4. FIG.

  In addition, since the image processing apparatus 1 can control the wiping cycle (waiting time or interval time) of the wiper 160, the cycle in which the image processing apparatus 1 acquires an image from the image sensor 30 can be changed according to the wiping cycle of the wiper 160. When the rainfall is low, the raindrop detection process can be executed with a long cycle, and when the rainfall is heavy, the raindrop detection process can be executed with a short cycle, thereby improving the efficiency of image processing.

  Furthermore, the brightness of the outside world by a street light or a car light may be measured based on the image input in FIG. 16, and the wiping cycle of the wiper 160 may be changed according to the brightness. For example, the wiping cycle is accelerated when it is bright, and is delayed when it is dark. Thereby, in the case of drizzle, it can suppress that a street light and a car light are scattered by raindrops, and a visual field becomes bad.

  Further, it is possible to control whether the fog lamp is turned on / off or the high beam is prohibited / permitted based on the calculated rainfall. For example, as shown in FIG. 21, when there is a lot of rain, the fog lamp may be automatically turned on to prohibit the high beam. In FIG. 21, when the fog lamp is not lit when there is a lot of rain, it automatically turns on or outputs an alarm prompting the lighting to the driver, and nothing is done when the fog lamp is lit. If the high beam is not lit, lighting is prohibited. If the high beam is lit, an automatic turn-off or a warning prompting the turn-off is output to the driver. Also, when the rainfall is low, the fog lamp is not turned on / off, and the high beam is automatically turned on / off according to the position (distance, direction) of the preceding vehicle or oncoming vehicle. Thereby, it can suppress that the light of a headlight is scattered by rain, and can improve a visual field.

  Next, when detecting cloudiness adhered to the inside of the window glass 4 due to the difference between the room temperature and the outside temperature, the image processing device 1 causes the second light source 3 to emit light and irradiate the window glass 4. When there is no fog on the inside of the window glass 4, the light from the second light source 3 is totally reflected inside the window glass 4 and enters the image sensor 30. When the window glass 4 is clouded, the light from the second light source 3 is scattered by the clouding, and a part of the light enters the image sensor 30 as reflected light. The storage element in the image processing apparatus 1 stores in advance the maximum brightness of reflected light when there is no fog inside the window glass 4 as an initial value. Then, the image signal from the image sensor 30 is input, the maximum brightness of the image is obtained, and the difference from the stored initial value is calculated by the image processor. If the difference calculated by the image processor is large, it is determined that fogging has occurred inside the window glass 4.

  FIG. 22 shows a flow of an example of the fog detection process. In step 11, the image processor turns on the second light source 3 to illuminate the window glass 4 and inputs an image signal from the image sensor 30. In step 12, the maximum brightness of the light reflected inside the window glass 4 is detected from the image signal. In step 13, the difference between the detected brightness and the initial value stored in the storage element is calculated. In step 14, the difference is converted to a cloudiness level to detect cloudiness, and the flow in FIG. 22 is terminated. When the image processing apparatus 1 detects clouding adhering to the inside of the window glass 4, the image processing apparatus 1 and the air conditioner are interlocked to remove the clouding. To control.

  Next, another embodiment of the present invention will be described. In the image processing system, the image acquired by the image processing apparatus 1 is processed to measure the brightness of the outside light, and based on the measured brightness of the outside light, for example, day and night are discriminated to determine the first light source 2 and the first light source 2. The first light source 2 and the second light source 3 may be controlled by the image processing apparatus 1 so that the position irradiated by the second light source 3 is changed.

  This will be specifically described with reference to schematic diagrams shown in FIGS. In this embodiment, the first light source 2 and the second light source 3 are composed of a plurality of light emitting elements. In the example of FIGS. 23 and 24, two first light sources 2-1 and 2-2 and two second light sources 3-1 and 3-2 are used.

  For example, since ambient light including sunlight enters the image sensor 30 during the daytime, a bright image can be obtained by capturing an image at a timing when the first light source and the second light source are turned off. Therefore, if the brightness of the image captured by the image sensor 30 is analyzed by the image processor, day and night can be determined. Since the image sensor 30 faces forward, the road surface on which the host vehicle travels is also included in the image. If it is daytime, the road surface is a dark area in the image captured by the image sensor 30. Therefore, if the image captured by the image sensor 30 includes a dark region, the light source may be selected so as to irradiate the dark region. The light source 2-2 is installed slightly below the light source 2-1 disposed at substantially the center position of the image sensor 30, and can irradiate the window glass 4 near the road surface of the image. The light source 3-2 is arranged closer to the window glass 4 than the light source 3-1, so that the window glass 4 near the road surface of the image can be irradiated. If it is daytime, the base voltage of the transistor 9-3 or the transistor 9-4 is controlled so that the light source 2-2 or the light source 3-2 emits light in the driving circuit shown in FIG. 2 or a light source 3-2 is caused to pass a predetermined current. Otherwise, a predetermined current is supplied to the light source 2-1 or the light source 3-1 by controlling the base voltage of the transistor 9-1 or the transistor 9-2 so that the light source 2-1 or the light source 3-1 emits light. Make it flow. As a result, it is possible to detect raindrops and cloudiness by reducing the influence of bright ambient light such as sunlight.

  In addition, the brightness of the external light is measured by processing the image acquired by the image processing apparatus 1, and the light emitted from the first light source 2 and the second light source 3 is based on the measured brightness of the external light. A control method of the image processing system that changes the strength is also conceivable. The intensity of the ambient light around the automobile can be determined by the brightness of the image in the same way as the above-described day / night determination. When the external light is bright, the intensity of light emission from the light source is increased. When the external light is dark, the intensity of light emission from the light source is decreased. This control method can also be performed using the configuration of FIG. For example, by controlling the base voltages of the transistors 9-1, 9-2, 9-3, and 9-4 by the image processor, currents that flow through the light sources 2-1, 2-2, 3-1, 3-2 And the intensity of light emission of the light sources 2-1, 2-2, 3-1, 3-2 can be changed. According to the image processing system using this control method, raindrops and cloudiness can be detected without being affected by outside light even in the daytime. Further, when infrared light is used for the light sources 2-1, 2-2, 3-1, 3-2, it is possible to detect raindrops and cloudiness at night without illusioning other vehicles or pedestrians.

  Further, in another embodiment of the present invention, as shown in FIG. 25A, the second light source 3 is constituted by a plurality of light sources arranged in an array. In the example shown in FIG. 25A, the second light source 3 is constituted by four light sources 3-11, 3-12, 3-13, and 3-14. FIG. 25C shows a lighting circuit configuration example of a plurality of light sources arranged in an array. In the circuit of FIG. 25 (c), the four light sources 3-11, 3-12, 3-13, 3-14 are controlled to light up simultaneously. Alternatively, the light source may be turned on using a plurality of transistors 9-21 to 9-24 shown in FIG. According to the circuit of FIG. 25 (d), the four light sources 3-11, 3-12, 3-13, 3-14 are controlled so as to be turned on one by one in order. If the light source of the present embodiment is used, not only the degree of fogging in the window glass 4 but also the clouding direction and clouding speed can be measured using an image processor or the like.

  For example, as shown in FIG. 25 (b), when the fog occurs on the inner upper part of the window glass 4, the brightness of the reflected light from the window glass 4 is the same as the reflected light from the light source 3-11 first. If the reflected light from the light source 3-12 becomes darker and then the reflected light becomes darker in the order of the light source 3-13 and then the light source 3-14, the fogging inside the window glass 4 is the light source. It represents that the light travels from 3-11 to the light source 3-14. If the reflected light from the window glass 4 becomes darker in the order of the light source 3-14 to the light source 3-11, the traveling direction of fogging inside the window glass 4 becomes the opposite direction to the above example.

  Further, the time interval until the reflected light from the light source 3-11 becomes dark and then the reflected light from the light source 3-12 becomes dark, the light source 3-12 becomes dark and then the light source 3-13 becomes dark. By calculating the time interval until, the traveling speed of fogging inside the window glass 4 can be calculated. In this case, since the interval between the light sources 3-11 to 3-14 is also used for the calculation, it must be stored in the storage element. According to the present embodiment, since the speed is known together with the information whether the cloudy area is developing from the periphery of the window glass 4 or whether the cloudy area is decreasing, the air volume from the defroster can be finely controlled, The fogging of the window glass 4 can be efficiently removed.

  When cloudiness is detected inside the window glass 4, first, the air conditioner may be controlled to remove the cloudiness, and then the wiper may be controlled based on the amount of rain detected by the raindrop detection process. Thereby, it can prevent that it becomes difficult to detect a raindrop by the cloudiness of the window glass 4, and the precision of the raindrop detection process result of FIG. 16 can be improved.

  By using the image processing system of this embodiment, it is possible to accurately detect raindrops and cloudiness on the window glass 4 and to finely control the wiper and the defroster, so that the driver has better visibility and detects lanes and vehicles ahead. It becomes easy to do. Therefore, the control content of the host vehicle can be changed based on the detected lane information and information on the preceding vehicle.

  Further, in the image processing system of the present embodiment, the focal length of the lens 10 is fixed at a long distance, but the focal length is changed by using a zoom lens or the like, and the focal point is far away from the position of the window glass 4 and the automobile. You may comprise so that an image may be acquired collectively. According to the image processing system having this configuration, a clear image can be obtained in the distance and the vicinity. Therefore, the detection of the tail lamp of the preceding vehicle and the head lamp of the oncoming vehicle that travels in the distance, and the raindrops attached to the window glass 4 Cloudiness can be detected by the same image processing system.

  It takes time to change the focal length using a zoom lens or the like. In order to reduce the time required for changing the focal length, the lens 10 provided in the image processing apparatus 1 may be a lens having an aperture ratio of a predetermined value or more (for example, F2.5 or more). According to the image processing system of this configuration, it is possible to reduce the blurring of the images in the distance and the vicinity even when focusing on the distance, so it is possible to detect the headlamps of the preceding vehicle traveling on the distance and the tail lamps of the oncoming vehicle. In addition, it is possible to easily detect raindrops and cloudiness adhering to the window glass 4 in the vicinity.

  In addition, when the image processing apparatus 1 processes an image and detects raindrops or cloudiness, the raindrops or cloudiness are determined according to the area where the image is processed, that is, the areas irradiated by the first light source 2 and the second light source 3. You may comprise so that the threshold value which detects may be changed. As shown in FIGS. 1 and 23, when the window glass 4 is inclined, the size of raindrops in the image acquired by the image sensor 30 and the intensity of the reflected light may be different. By changing, raindrops and cloudiness can be detected well.

  Furthermore, in a vehicle equipped with the image processing system of the present invention, the wiper and the defroster can be automatically controlled using the foreign object detection result of the window glass of the image processing system. Further, in the automobile equipped with the image processing system of the present invention, the foreign matter on the window glass can be removed, so that the front visibility is improved and the lane and the front vehicle are easily detected. The controllability of the vehicle is improved, such as distance control and lane departure warning.

  In the region of the window glass 4, the infrared light is transmitted to the region in which the reflected light of the light emitted from either the first light source 2 and / or the second light source 3 is imaged by the image sensor 30. You may make it give the characteristic to do. In recent years, the window glass 4 of automobiles has been increasingly subjected to infrared cut coating in order to reduce unpleasant elements from driver's solar radiation. In this case, since infrared wavelengths from the first light source 2 and the second light source 3 are absorbed, sufficient reflected light cannot be obtained, and raindrop detection performance deteriorates. Therefore, as shown in FIG. 26, the infrared light is applied to the area 40 where the reflected light of the light emitted from the first light source 2 and / or the second light source 3 is imaged by the image sensor 30. If the infrared cut coating is not applied or is removed, for example, the reflected light can be obtained efficiently and the raindrop detection performance can be improved.

  In addition, as shown in FIG. 27, the processing is performed so that the region 50 in which the reflected light of the light emitted from the second light source 3 is captured by the image sensor 30 in the region of the window glass 4 is likely to be clouded. You may give it. For example, the area 50 is made of frosted glass or is easily fogged by applying a water repellent coat, etc., so that the occurrence of fogging on the entire surface of the window glass 4 can be detected quickly, and before the driver's view is obstructed, Will be able to control.

  The focal point of the lens 10 installed in the image processing apparatus 1 may be between infinity and the window glass 4. As described above, in order to detect the raindrop 8, rather than the focus of the lens 10 being on the raindrop 8 on the surface of the window glass 4, the recognition rate as a circle is higher when a slight blur occurs, Raindrop detection performance is improved. Further, when the lens 10 is in focus at infinity, when detecting the tail lamp of a preceding vehicle traveling far away, particularly when the image sensor 30 is composed of a complementary color filter with a small number of pixels in color, the image sensor If the size of the tail lamp on 30 is about 1 to 4 pixels, color reproduction may not be performed and red may not be recognized, making it difficult to accurately detect the tail lamp of the preceding vehicle. By focusing the lens 10 closer to infinity than the infinity, the tail lamp of the preceding vehicle is blurred and has a size of 4 pixels or more, and the color can be accurately reproduced in the image sensor 30. This makes it possible to fix the focus of the lens 10 that simultaneously satisfies the detection of raindrops and the detection of the tail lamp of the preceding vehicle.

  Further, according to the present invention, it is possible to detect a foreign matter such as raindrops or cloudiness adhering to the windshield, and at the same time, to detect a tail lamp of a preceding vehicle.

  The present invention is not limited to the automobile shown in the above embodiment, but also drops such as raindrops attached to the surface of various types of window glass such as glass used for railway vehicles, ships, aircraft, etc. or window glass of general buildings, and It can also be used to detect cloudiness and dust.

The schematic diagram of an example of the image processing system by this invention. The figure which shows the example of the raindrop and cloudy image image | photographed with the image pick-up element. The schematic diagram of the other example of the image processing system by this invention. The schematic diagram of the other example of the image processing system by this invention. Explanatory drawing of the optical characteristic of an optical filter. It is a schematic diagram of an optical filter. The figure explaining an example of a light distribution control function. The figure which shows the characteristic of the infrared-light cut of a filter. The figure which shows an example of a raindrop detection function. The figure which shows the structural example of an optical filter. The figure which shows the structural example of an imaging part. Explanatory drawing of the imaged image. The figure for demonstrating the effect acquired by inclining an imaging part or an imaging device. The figure for demonstrating the area | region required when performing lane detection. The figure which shows an example of operation | movement of the light emission drive circuit of a light source, and an image processing apparatus. The flowchart which shows the procedure of a raindrop detection process. The figure for demonstrating a wiper and washer control. The flowchart of a wiper control process. The figure explaining the example of the wiper starting method. The figure explaining the other step used for a wiper control process. The figure explaining the example of the control method of a fog lamp and a high beam. The flowchart which shows the procedure of a cloudiness detection process. The schematic diagram of the other example of the image processing system by this invention. The figure which shows the example of the light emission intensity control circuit of a light source. The schematic diagram which shows the 2nd light source which consists of a several light-emitting body, and its usage example. The schematic diagram of the other example of the image processing system by this invention. The schematic diagram of the other example of the image processing system by this invention.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Image processing apparatus, 2 ... 1st light source, 3 ... 2nd light source, 4 ... Window glass, 5 ... Case, 6 ... Dashboard, 7 ... Heat conduction member, 8 ... Raindrop, 9-1-9- DESCRIPTION OF SYMBOLS 4 ... Transistor, 10 ... Lens, 11 ... Passage, 20 ... Optical filter, 30 ... Image sensor, 40 ... Raindrop detection area, 50 ... Cloudiness detection area, 210 ... Preceding car, 220 ... Oncoming car, 51a, 51b, 51c, 51d ... Infrared cut region, 52a, 52b, 52c, 52d ... Infrared bandpass region, 81 ... Straight line, 82 ... Curve, 100 ... Wiper / washer control unit, 150 ... Motor, 160 ... Wiper, 170 ... Washer control unit

Claims (37)

A first light source that emits light toward the glass;
An image sensor that images light from the first light source reflected by the foreign matter attached to the glass;
An optical filter provided between the glass and the imaging device and transmitting a specific wavelength of the reflected light;
An image processing apparatus for processing an image captured by the image sensor;
With
An image processing system comprising a lens in front of the image sensor, wherein the focal point of the lens is set farther than the position of the glass.   The image processing system according to claim 1, wherein the imaging element has a region that receives light through the optical filter and a region that receives external light that has passed through the glass without passing through the optical filter. An image processing system.   The image processing system according to claim 1, further comprising an infrared cut filter between the optical filter and the glass.   The image processing system according to claim 1, wherein the glass is a windshield of an automobile, and includes a second light source that irradiates light toward the glass, and the imaging element is reflected by a foreign substance attached to the glass. The light from the first light source is imaged and the light from the second light source reflected by the foreign matter attached to the glass is imaged, and the optical filter is on the longer wavelength side than the visible region and detected by the image sensor. An image processing system that transmits light in a wavelength region to be transmitted.   5. The image processing system according to claim 4, wherein the second light source is provided on a dashboard in a driver's cab of an automobile.   5. The image processing system according to claim 4, wherein the image processing device detects raindrops attached to an outer surface of the glass by light emitted from the first light source, and uses light emitted from the second light source. An image processing system for detecting cloudiness adhering to the inner surface of the glass.   5. The image processing system according to claim 4, wherein the image processing device, the first light source, and the second light source are disposed in a case, and an air circulation passage is provided between the case and the glass. An image processing system characterized by that.   5. The image processing system according to claim 4, wherein a tail lamp of a preceding vehicle is detected together with the detection of the foreign matter adhering to the glass.   5. The image processing system according to claim 4, wherein the optical characteristics of the optical filter differ depending on the area of the optical filter.   The image processing system according to claim 9, wherein a part of the optical filter is an infrared cut area for cutting infrared light, and at least a part other than the infrared cut area has a transmittance peak in the infrared area. An image processing system characterized by an infrared bandpass region.   The image processing system according to claim 10, wherein the optical filter has the infrared bandpass region on an upper side in a state where the optical filter is installed between the glass and the imaging device.   The image processing system according to claim 10, wherein the optical filter has the infrared bandpass region on an upper side and the infrared cut region on a lower side in a state of being installed between the glass and the imaging element. An image processing system comprising:   The image processing system according to claim 10, further comprising: a lens in front of the image sensor, wherein the lens is coated with a coating having the same optical characteristics as the optical filter.   11. The image processing system according to claim 10, wherein a coating having the same optical characteristics as the optical filter is applied to a cover portion provided in front of the image sensor.   The image processing system according to claim 10, wherein the image processing device performs raindrop detection by processing an image captured in the infrared bandpass region.   The image processing system according to claim 10, wherein the image processing apparatus detects a preceding vehicle by processing an image captured in the infrared cut region.   The image processing system according to claim 10, wherein the image processing apparatus performs lane detection by processing an image captured in the infrared cut region.   The image processing system according to claim 4, wherein a field of view of the image sensor faces downward from a horizontal direction.   5. The image processing system according to claim 4, wherein the image processing device detects a rain amount by processing an image picked up by the image sensor, and controls a wiper or a washer of an automobile based on the detected rain amount. Image processing system.   20. The image processing system according to claim 19, wherein control of the wiper is prohibited when the shift lever of the automobile detects a parking state.   The image processing system according to claim 19, wherein a rain determination threshold value for controlling the wiper is changed according to a traveling speed of the automobile.   20. The image processing system according to claim 19, wherein after the wiper is activated, control of the wiper based on the detected rainfall is started after a predetermined waiting time has elapsed.   23. The image processing system according to claim 22, wherein the length of the waiting time is controlled using at least one of a traveling speed of a vehicle and a raindrop diameter.   The image processing system according to claim 19, wherein the image processing device detects fogging adhered to the inner surface of the glass by reflected light of light emitted from the second light source, and the fogging adhesion occurs on the inner side of the glass. An image processing system that controls a wiper or a washer based on the detected rainfall when there is no rain.   20. The image processing system according to claim 19, wherein a torque of a motor that drives the wiper is detected to determine a water repellent state of the glass or a damaged state of the blade of the wiper.   The image processing system according to claim 19, wherein fog lamp lighting control is performed based on the detected rainfall.   20. The image processing system according to claim 19, wherein the high beam lighting of the headlamp is controlled based on the detected rainfall.   20. The image processing system according to claim 19, wherein the brightness of external light is measured, and the operation speed of the wiper is changed according to the measured brightness.   The image processing system according to claim 19, wherein the intermittent time of the wiper is controlled using at least one of the detected rainfall and rain particle size.   The image processing system according to claim 19, wherein an intermittent time of the wiper is controlled in accordance with a traveling speed of the automobile.   5. The image processing system according to claim 4, wherein the reflection luminance from the glass is measured, and dirt and / or raindrops on the glass are determined using the measured luminance.   32. The image processing system according to claim 31, wherein the wiper is driven when dirt is detected, and the washer is controlled when dirt is further detected at the same position after the wiper is driven.   5. The image processing system according to claim 4, wherein the rain adhesion state of the glass is detected to determine the water repellent state of the glass or the damage state of the blade of the wiper.   5. The image processing system according to claim 4, further comprising a mirror, wherein the light emitted from the second light source is reflected by the glass, then reflected by the mirror, and further reflected by the glass and incident on the imaging device. An image processing system. The image processing system according to claim 4.
The second light source is a plurality of light emitters arranged in an array,
The imaging element images light from the second light source reflected by fogging attached to the glass,
The image processing device processes the image and measures at least one of the degree of adhesion of the fogging to the glass, the traveling direction of the fogging, and the adhesion speed of the fogging to the glass. Image processing system.   The image processing system according to claim 1, wherein the optical filter transmits light in a wavelength region in which the transmittance of the glass is equal to or higher than a predetermined value and the sensitivity of the imaging device is high. 37. The image processing system according to claim 36 , wherein the optical filter transmits wavelengths in the visible light region.

Images